Qing Chen Email & Phone Number

US

Payment & Fraud Detection

Washington, District Of Columbia, US

• Developed and implemented a tree-based credit scoring machine learning model for PAMF’s nano loan product in Madagascar, disbursed more than 2 million loans for over 300 thousand customers. Reduced default rate by more.
• Built end-to-end machine learning pipeline to generate reproducible ML experiments to track the full ML lifecycle of feature processing, model training, model evaluation and model deployment. Improved team’s modeling.
• Built ETL pipeline on dataiku to validate, clean and integrate multiple data sources. Processed cleaned data to create insights to monitor model performance.
• Implemented the core data science module (including transaction tagging, summary feature calculation, and credit score modeling) of Lokyata BankAnalyze to analyze bank transaction data to provide credit worthiness.
• Implemented the core functions of Lokyata self-service business lending decision platform, including variable builder, rules builder, scorecard builder, decision matrix builder, and AI model builder. Deployed functions.

Washington, District Of Columbia, US

• Researched, developed and implemented machine learning models/algorithms (including various tree-based ensemble models and neural networks) to build Lokyata's core credit scoring engine.
• Built data science pipeline to process multiple data sources, including data cleaning, statistical analysis, feature engineering and model prototyping.
• Trained and recalibrated machine learning models, ran various machine learning tests, experiments and impact simulations to fine tune models.
• Designed machine learning systems for AWS cloud deployment.

Washington, D.C., US

• Researched correlation between mouse behavior (motion and social behavior) and neuron firing in mouse brain.
• Researched and applied unsupervised learning methods (including PCA, NMF) to extract latent features from neurons firing in mouse brain.
• Implemented deep learning models (including CNN, LSTM) for seizure early prediction, using phase transition theory for feature extraction.

• Researched various text mining methods (including LDA topic modeling and Glove) to build a recommender system for Zoom Social app.
• Built a extensible system that categorized location-based Twitter content into 22 general categories using customized CNN with Tensorflow for Zoom Social app, achieved 0.85 F1-macro for offline evaluation.
• Designed a big data platform, and built a model pipeline using LSTM with NLP features (Glove word embedding and sentiment analysis) for location-based sales forecasting in a supply chain & logistics big data.

Washington, D.C., US

• Helped students understand basic machine learning methods: logistic regression, SVM, Bayesian network, perceptrons, Boltzmann machine, etc.
• Implemented various deep learning models (DNN, CNN, Multitasking) with Tensorflow for drug target predictioin, increased accuracy by 30%. (https://github.com/qchen7/ML_Project_Drug_Design)

• Researched and analyzed the impacts of artificial intelligence and automation on US labor market based on data from Bureau of Labor Statistics (using R and SQL), visualized results using Tableau. Presented results at.

Fairfax, Virginia, US

• Researched the underlying problems and parameter choices of Lattice-based and Multivariate Cryptography.

Fairfax, Virginia, US

• Researched various EM algorithms and models of network topology to solve temporal sensing and resource allocation problems.
• Helped integrate the sensor code with spectrum sensing algorithms based on Hidden Bivariate Markov Model (HBMM) using Python and C++.

• Helped organize student events and communication with other departments
• Maintained US-China dual degree program course database

Master Of Science - Ms, Data Science

Bachelor Of Science - Bs, Ece